Identification of biomarkers associated with metabolic cardiovascular disease using mRNA-SNP-miRNA regulatory network analysis

It is estimated that in 2019 463 million adults aged 20–79 years would develop diabetes with type 2 diabetes (T2DM) accounting for > 90% of cases [1]. The incidence of T2DM has greatly increased in recent years and has become a great threat to human health worldwide. T2DM is one of major risk factors of cardiovascular diseases (CVD), but its mechanism of action is not fully understood [2]. Even newly diagnosed diabetics were reported to have at least one vascular complication, and CVD is the leading cause of death in T2DM patients [3, 4]. Furthermore, people with T2DM are 2–6 times more likely to die of CVD than those without diabetes [5]. However, few biomarkers have been identified to diagnose CVD at the early stage of T2DM.

MicroRNAs (miRNAs) bind to complementary sequences of their target mRNA by base pairing to induce mRNA degradation and/or inhibit translation, thus affecting gene expression after transcription [6]. MiRNA has been found to have high specificity for disease status and may be used as a potential biomarker to predict disease progression. In recent years, multiple miRNAs have been demonstrated to be involved in angiogenesis and endothelial cells dysfunction [7‐9]. Expression of miRNAs can be affected by many factors. The miRNA seed region can specifically bind to miRNA recognition element (MRE) in the 3′UTR (3′ untranslated region) of their target mRNA [10]. Therefore, any anomalies of miRNA-MRE interaction may have an impact on gene expression and lead to diseases [11‐13]. Single nucleotide polymorphism (SNP), the most common genetic variants, can interfere with the base pairing between miRNA and its target mRNA, affecting normal expression of genes and eventually contributing to disease pathogenesis [14]. Genome-Wide Association Studies (GWAS) have found that genetic variants play an important role in pathogenesis of CVD [15], and association between some genetic variants and CVD has critical biological significance for T2DM patients [16, 17]. Genome-wide expression quantitative trait locus (eQTL) analysis is an effective method to study the effect of SNPs on gene expression [18, 19]. However, eQTL analysis mainly focuses on mRNA expression and rarely involves miRNA, which leads to incomplete interaction patterns [20].

To more comprehensively reveal the complicated association between mRNA, miRNA and SNP, and to find potential biomarkers with high specificity and sensitivity to diagnose CVD at the early stage of T2DM, we performed an integrative analysis. EQTL analysis and gene ontology (GO), KEGG pathway enrichment analyses were also conducted to better understand the connection between mRNA, miRNA and SNP, and their potential effect on CVD in T2DM patients.

CVD is the most common cause of death among patients with T2DM, imposing a heavy burden on the economy of individuals and societies. Some studies have investigated the underlying mechanism of CVD and T2DM, and found that miRNA and SNP play important role in the occurrence and development. However, there is still a lack of biomarkers with good sensitivity and specificity for early detection and diagnosis of CVD in T2DM patients. Constructing the mRNA-SNP-miRNA interaction network can help to reveal the underlying correlation between genetic variants and diseases. But there is a lack of miRNA sequencing data of CVD in T2DM patients, leading to few evidence and certain limitations. We combined mRNA, miRNA and SNP data by searching GEO database and conducting miRNA sequencing. Our present study identified 16 mRNA-SNP-miRNA interaction trios, which revealed that SNPs in the binding region between miRNA and mRNA can interfere gene expression. Previous studies have proved that miRNAs play an essential role in the occurrence and development of various diseases [21], and our findings may increase understanding of the underlying mechanism.

In this study, we identified 56 differentially expressed genes and 16 differentially expressed miRNAs between T2DM patients and T2DM patients with CVD. According to mRNA-miRNA co-expression network, let-7i-5p and miR-320c were both related to 6 genes. Previous study showed that let-7i-5p was involved in the regulation of cardiac cell cycle, and inhibition of let-7i-5p may be a potential strategy for cardiac repair after ischemic injury [22]. In addition, downregulated let-7i was observed in dilated cardiomyopathy and low expression of let-7i was associated with poor clinical outcomes of patients with dilated cardiomyopathy [23]. Downregulated let-7i-5p was observed in cardiomyocytes during hypoxia injury and let-7i-5p pathway was used to suppress hypoxia-induced apoptosis and mitochondrial energy metabolism dysfunction in vitro [24]. Taken together, let-7i-5p was demonstrated to be associated with the pathogenesis of CVD, and may play an important role in the pathophysiological process of CVD. Interestingly, the six genes associated with downregulated miR-320 were up-regulated which indicated that miR-320 may bind to complementary sequences of mRNA of these genes to induce mRNA degradation or inhibit translation, thus affecting gene expression.

In our present study, KEGG pathway enrichment analysis indicated significant enrichment in pathways including B cell receptor signaling pathway and hematopoietic cell lineage. Previous studies suggested that compared with healthy people, CD19 was related to B cell receptor and significantly downregulated in patients with acute myocardial infarction [25]. However, this study showed that CD19 associated with B cell receptor signaling pathway was upregulated in T2DM patients with CVD, which could attribute to the mechanism of CVD in T2DM patients. Moreover, mutations in the TET2 that can promote clonal hematopoiesis were associated with an increased risk of atherosclerosis [26]. Interestingly, we found that TET2 was target gene of 7 differentially expressed miRNAs (miR-133a-3p, miR-6817-3p, miR-873-5p, miR-581, miR-210-3p, miR-202-5p, miR-2355-3p, let-7i-5p, miR-196a-5p, miR-760) in our study which highlighted that the hematopoietic cell lineage and differentially expressed miRNAs may be involved in the progression of CVD in T2DM patients. In addition, KEGG pathway enrichment analysis showed PI3K-Akt signaling pathway was associated with PDGFRB, CD19 and TCL1A. Some studies confirmed that PI3K-Akt signaling pathway plays an important role in the pathophysiology of vascular diseases [27]. Activated PI3K-Akt signaling pathway has been proven to improve insulin sensitivity, regulate glucose and lipid metabolism, and protect vascular endothelial cells [28]. In addition, activating IRS/PI3K/Akt pathway activity may play an anti-atherosclerotic role [29]. Interestingly, KEGG pathway enrichment analysis indicated that linoleic acid metabolism was related to ALOX15, and a growing number of studies reported that linoleic acid was associated with the prevention of T2DM and CVD [30, 31]. Therefore, we can reasonably speculate that PI3K-Akt signaling pathway and linoleic acid metabolism may play a role in the pathogenesis of T2DM complicated with CVD.

MRNA-SNP-miRNA interaction network included 16 SNPs, 9 mRNA and 11 miRNAs, but only miR-581, rs325009 and CEP41 were related to each other. Previous studies suggested that CEP41 was a new regulator for angiogenesis which promoted angiogenesis through the HIF1A-Aurka-VEGF pathway and was involved in endothelial cell migration which was involved in the pathogenesis of CVD [32]. Our study showed that RASGRP3 was associated with 4 SNPs in the mRNA-SNP-miRNA interaction network, and each intermediated one pairs of miRNA-mRNA correlations, which indicated that RASGRP3 may play a key role in the underlying mechanism. Paramjeet and colleagues observed that RASGRP3 could affect the role of endothelial cells in angiogenesis in diabetic mice by mediating endothelial cell signal transduction [33]. We further explored functions of the rest of mRNA in the mRNA-SNP-miRNA interaction network, and found that KRT1 might be associated with CVD. Gao et al. found that inhibiting KRT1 can activate Notch signaling pathway, thereby inhibiting the inflammatory response and endoplasmic reticulum stress of vascular endothelial cells in coronary atherosclerosis [34]. Furthermore, it has been reported that inhibition of KRT1 expression can improve myocardial ischemia–reperfusion injury by activating Notch signaling pathway [35].

Although previous studies have shown that miRNA-related SNPs play a role in the pathogenesis of T2DM complicated with CVD, there is a lack of relevant researches that combine mRNA, miRNA and SNP for analysis. We searched GEO database and found that there are few studies involving miRNA sequencing for T2DM patients and T2DM patients with CVD. Hence, we recruited six diabetes patients and five diabetes patients with ischemic heart disease from communities in Beijing to obtain their miRNA expression profile. Previous studies suggested that miR-196a-5p and miR-202-5p may play a role in the pathogenesis of diabetes and its cardiovascular complications. Omer et al. revealed that compared with normal group, miR-196a-5p was downregulated in CVD group without statistically significance, and they indicated that miR-196a-5p may be a potential biomarker for the diagnosis of coronary artery disease and acute coronary syndrome [36]. However, our present study suggested that miR-196a-5p was differentially upregulated in CVD group, which was inconsistent with previous study. We attributed the difference to the limited sample size, and studies with large samples are needed to verify the results. In addition, up-regulated miR-202-5p was found to have a protective effect on the heart of mice with myocardial ischemia–reperfusion injury [37]. But our present study identified that miR-202-5p was differentially upregulated in CVD group, and we considered that it was due to the different species in two studies which weakened the comparability.

We successfully constructed the mRNA-SNP-miRNA interaction network to visualize the relationships between mRNAs, SNP and miRNAs. However, there was no statistical significance in Pearson correlation analysis of mRNA and miRNA connected by overlapping SNPs in the mRNA-SNP-miRNA interaction network. Considering this study, the reasons may be associated with limitations in this study. First, the sample size of miRNA is so small that its results are not representative enough, and future studies including larger samples are required; second, mRNA and miRNA expression data were obtained from different samples, resulting in bias although we searched eligible databases by defining the CVD clearly and setting covariates to eliminate the bias, and miRNA, mRNA and SNP data from same experiment are needed to provide more convincing research results. Moreover, some miRNAs cannot inhibit mRNA expression but can inhibit protein translation. Hence, the co-expression analysis of mRNA and miRNA was not statistically significant, which did not indicate that there was no regulatory relationship between them. Further study on the significance of miRNA and mRNA at protein level is required.

We demonstrated mRNA-SNP-miRNA interaction network and identified important pathways contributing to metabolic cardiovascular disease. These results suggested that downregulated let-7i-5p, and upregulated RASGRP3, KRT1 and CEP41 may play crucial roles in molecular mechanisms underlying the initiation and development of CVD in T2DM patients. However, further studies regarding the role of these genes and miRNAs in progression of CVD in T2DM patients are required.